Download Cisco Tidal Enterprise Scheduler 6.2: Unix User Guide - Scheduler Concepts and more Slides Operating Systems in PDF only on Docsity!
C H A P T E R
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
Scheduler Concepts
Overview
Welcome to Enterprise Scheduler! This tutorial will introduce and guide you through the features and functions of the world’s premier network scheduling solution, Enteprise Scheduler. Each chapter in this guide builds upon the foundation of previous lessons. If this is your first time working with production scheduling software, or your first time using Enteprise Scheduler, you should work through each chapter consecutively. More experienced users might want to move directly to the lesson in which they are interested; however, many exercises utilize the results of previous exercises. As you progress through the tutorial, more advanced and detailed topics are covered. By the end of this manual, you will be familiar with most of the major features of Enteprise Scheduler, and you will have the knowledge to create and maintain your own production schedules.
Prerequisites
Before using this tutorial, you must copy the sample files that are used in the tutorial to your machine. These files are in a .tar file on the Scheduler installation DVD. To copy the files to your machine, follow these steps:
Step 1 Install the agent as root as described in the Installation and Configuration Guide before installing the tutorial sample files. Step 2 The .tar file with the tutorial sample files is found on the DVD at \agent\Unix\tutorial.tar. Copy this file onto the agent machine to the / (root level) directory of the Unix agent you will run jobs from. Step 3 Change the directory permissions: chmod 777 * Step 4 Unpack the .tar file with the following command: tar -xvf tutorial.tar The following sample files should be unpacked:
- UNIX_log.txt - UNIX_Log2.txt - UNIX_TEST_1.sh - UNIX_TEST_2.sh - UNIX_TEST_3.sh
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
System Configurations
- UNIX_TEST_4.sh - UNIX_TEST_5.sh - UNIX_TEST_JEVENT.sh Step 5 Change the permissions for the unpacked tutorial sample files with the following command: chmod 777 * Step 6 Change the ownership of the files to group: **chgrp -R ***
Note You can identify which group you are a member of by entering: id This returns the name of the group you belong to.
Step 7 Change the ownership of the files to a owner: **chown -R ***
Note The files can now be used with the lessons provided in this tutorial. To complete the exercises in this tutorial, you need to: Install Enteprise Scheduler in the default directory Scheduler (or the examples in this tutorial will not work properly. Select the Super User option in your User definition. Configure a default agent Create and have available the work day calendar
System Configurations
Enteprise Scheduler is used to schedule and manage jobs on several different systems — even different operating systems — through a single operator station. This single point-of-control means that you have only one Tidal Web client running on your PC to manage jobs across multiple systems. Masters are the central point for scheduling jobs on associated agents. Enteprise Scheduler jobs can only run on licensed agents.
Masters
The master is the UNIX system on which you install the “brains” of your Enteprise Scheduler network. You interact with the master using the Tidal Web client. The master performs the requested service on a local or remote agent, and then returns updated information to the Tidal Web client which displays the results. The master launches jobs on machines licensed as agents. An agent is software on another machine in the same network as the master, and runs jobs on behalf of the master. The Tidal Web clilent does not connect directly to the agent, but schedules and manages its jobs through the master.
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
Job Hierarchy
Job Hierarchy
Jobs are built on a hierarchy of job and job group ownership. A job group is a container for a set of jobs, usually part of a common application or department. The job group has its own name and set of runtime instructions. You can use job groups to submit jobs that either depend on each other, or should run together. For example, all the jobs in payroll can belong to a group called Payroll. The job group can provide default settings to all the child jobs that belong to it. Jobs and job groups are displayed in the Jobs pane. Job groups can save you the time it takes to set up job definitions because each job in the job group can inherit the characteristics of that job group. When you want to create several jobs with similar scheduling characteristics, you can define those jobs within a job group and set the scheduling characteristics in the job group definition. It is also possible to change scheduling characteristics at the job level even though the job belongs to a group. For example, if a job group is defined to run every Friday, then every job in that job group is automatically defined to run on Friday. If one job in the job group must run on Saturday, then that one job can be changed to the proper run day without affecting the other jobs — as long as you disinherit the job group calendar and change the calendar from within that job. The ultimate ownership of a job or job group belongs to either the user or a workgroup. A workgroup is a collection of users who can share access to the same jobs. Workgroups are displayed in the Workgroups pane.
Dependencies
Dependencies are prerequisite conditions that must be met before a job can run.
Date and Time Dependencies
The most common dependency is the date and time when Enteprise Scheduler executes a job. For example, you can schedule a job to run every Tuesday after 6:00 pm, except on holidays when it is not to run. Date dependencies are built using calendars. Time dependencies are specified within a job’s definition.
Job Dependencies
Jobs can also depend on other jobs reaching a particular status. For example, you can run Job51 after Job101 and Job207 have reached the status of Completed Normally. During the job’s life cycle, Enteprise Scheduler recognizes the current status of a job, such as: Table 1-1 Job Status
Status Description Waiting on Dependencies The job is waiting on Date, Time, Job, and/or File dependencies. Waiting on Resources The job is waiting for an execution slot. All Dependencies have been met. Waiting on Operator All the job’s dependencies are met and the job is waiting for the operator to release it.
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
Job Hierarchy
File Dependencies
A job can also depend on the status of a file. The state, size, creation or modification date of the file can all be taken into consideration. For example, you can run Job101 if the Unix file /payroll/data/trandata :
- Has been modified in the last twelve hours - Has a file size greater than 1024KB
Variable Dependencies
A job can also depend on the value of a user-defined variable. Enteprise Scheduler has a repository of user-defined variables that can be updated or incremented either manually or through an action associated with a job event or system event. For example, you can set a job to run when:
- Variable Printer Online is set to Yes. The Printer Online variable could be set by a job that changes printer settings and then issues an action changing the variable from No to Yes. - Variable Payroll Jobs is incremented to 15 by another job that increments the variable each time it runs.
Calendars
Calendars are used to determine what days to run jobs. Calendars let you schedule jobs to run on a periodic yet intelligent basis. For example, Labor Day in the United States is celebrated on the first Monday in September which falls on a different date each year. By defining Labor Day as the first Monday in September, you avoid the need to manually redefine it every year. You can also define calendar groups which combine individual calendars. For example, the 1st Half Holidays calendar group can include the New Year’s Day , President’s Day , and Memorial Day calendars.
Job Instances
A job instance is a specific, scheduled run of a job definition (job) by Enteprise Scheduler. One job can create many instances.
Active The job is actively running in the Production Schedule. Completed Normally The job completed normally. Completed Abnormally The job completed abnormally. Error Occurred An internal error occurred which prevented the job from running.
Table 1-1 Job Status
Status Description
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
Job Hierarchy
- Send email messages - Control a job instance in the Job Activity window - Alert an operator to a job condition - Send SNMP messages - Launch an unscheduled job (new job action) - Issue a log message - Update a user defined variable For example, you can define a job event that is triggered every time a job is cancelled by an operator. When a job cancellation occurs, you can have an email sent to you and a message sent to your SNMP management software noting this event. A system event operates identically to a job event , except that the master originates the event rather than a job. System events define global conditions versus a job event defining conditions that affect jobs. For example, if an agent shuts down, a system event can be triggered to notify users of the problem. An email event is the detection of a specified text string in an email that arrives at a designated email account on an designated Exchange server. An email monitor is created to watch for the specified email. A file event is the detection of a file on an agent reaching a specified state. A file monitor is created to watch for a file the matches the specified conditions. A variable event is the detection of a variable reaching a specified value, whether the variable is on a local or remote master. A variable monitor is created to watch for the variable to reach the desired value.
Queues
Queues let you optimize throughput and allocate system resources for scheduled and unscheduled jobs. The Enteprise Scheduler queue manager assigns jobs to queues when all their dependencies have been met, and decides when to launch jobs based upon the available system resource slots. The maximum number of slots available is determined either by the limit that you set in the system queue, the sum of each queue’s limit or the sum of each licensed agent’s job limit. Queues can limit the number of jobs running on a computer or a network of computers at a given time.
- If the system is not running at its capacity, a job can run immediately provided that all of its dependencies are met. - If the system is running at its capacity, the Enteprise Scheduler Queue Manager decides which jobs launch based on a priority structure that includes the following in order of importance:
Table 1-2 Queue Priorities
Status Description Queue priority levels Jobs in active and open queues at higher priority levels run first. Queue limits Only jobs in queues not running at their allowable limit can be launched. Agent job limits Only jobs assigned to agents not running at their allowable limit can be launched. Job priority levels Jobs with the highest priority (assigned in the job definition) in the queue are run first.
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
Job Hierarchy
- Queues are displayed in a hierarchy. Each item in the hierarchy is a queue and can contain jobs. You define the queue limit to set the number of jobs that can launch from any individual queue. You also define a priority for each queue.
Queue Filters
Jobs are directed to a queue based on the queue filters that you define. These filters describe the job properties that must exist for the queue manager to assign a job to a particular queue. Some examples of the queue filters that direct jobs to queues are:
- Job class - Job name - Job owner - Job estimated runtime
Agent Lists
Enteprise Scheduler extends its capability for automatic job management through agent lists. An agent list describes a set of nodes on your network available to run jobs. Agent lists designate nodes as primary or alternate nodes for job submission, and allow jobs to be broadcast across all available nodes. Workload balancing algorithms can distribute jobs evenly among all available nodes.
Security Policies
Security policies restrict access to certain Enteprise Scheduler functions. The defined access rights can be saved as a security policy, and then assigned to one user or multiple users. For example, there might be different sets of users who:
- Administer Enteprise Scheduler - Create and schedule jobs for themselves and others - Operate the job schedule Using security policies, the users that create and schedule jobs can be restricted from modifying the schedules. Likewise, the operators can be restricted from creating jobs. Enteprise Scheduler includes default security policy templates that can be modified to create your own security policies. Each user within the supplied working model has a defined set of Enteprise Scheduler functions. When all the default security policies are in use, all aspects of scheduling are covered and available. The following table lists the system features available for each of the default security template: Table 1-3 Scheduler Security Policies
Default Security Policy Available System Features Scheduler_Admin The default for new installations. This includes all available functions. Administrator Configures users. User Creates, edits, and submits jobs. Creates workgroups and user-defined variables.
Cisco Tidal Enterprise Scheduler 6.2 Tutorial Guide for Unix OL-32204-
Job Hierarchy
